P&z, 42 (1990) 153-165 Elsevier

153

PAIN 01618

Psychosocial correlates of temporomandibular joint pain and dysfunction

Robert F. Schnurr a, Ralph I. Brooke ’ and Gary B. Rollman a

a Department of Psychoiogv and ’ Faculty of Dentistry University of Western Ontario, London, Ont. N&4 5C2 (Canada)

(Received 6 September 1988, revision received 13 February 1990, accepted 16 February 1990)

s-w This study examines psychological differences between temporomandibular joint pain and dysfunction (TMJPD) patients, pain controls, and healthy controls. Two hundred and two patients were classified, according to the diagnostic criteria of Eversole and Machado, as either myogenic facial pain (n = 42), internal derangement type I (n = 69), internal derangement type II (n = 85), or internal derangement type III (n = 6). Patients completed the Basic Personality Inventory, the Illness Behavior Questionnaire, the Multidimensional Health Locus of Control, the Perceived Stress Scale, and the Ways of Coping Checklist. Subjects also answered questions pertaining to TMJPD symptomatology, including chronicity and severity. After conservative treatment with simple jaw exercise and uhrasound, patients were contacted again at 5 months to complete follow-up questionnaires similar to those previously completed. Comparison groups were comprised of 79 patients attending outpatient physiotherapy clinics for pain-related injuries not involving the temporomandibular joint and 71 pain-free, healthy students. Data were analyzed using multivariate statistics. The results indicate a significant relationship between pain intensity (and to some extent chronicity) and diverse measures of personality among the pain controls but not among the TMJPD patients. This calls into question the validity of assuming individual pain disorders are subsets of a larger, homogen~us pain disorder population. TMJPD patients and pain controls score higher on hypochondriasis and anxiety than the pain-free controls but these elevations are not clinically significant. The elevations decrease to normal levels in response to a positive treatment outcome. There were no differences between the TMJPD patients and the pain controls on any of the measures. These results suggest that TMJPD patients do not appear to be significantly different from other pain patients or healthy controls in personality type, response to illness, attitudes towards health care, or ways of coping with SW%.%.

Key words: Pain; Personality; Temporomandibular joint pain and dysfunction

Introduction

Temporomandib~ar joint pain and dysfunc- tion (TMJPD), also commonly referred to as myofascial pain dysfunction syndrome, ‘is a con- dition in which pain, clicking of the tem- poromandibular joint, and limitation or deviation of jaw opening occur in association with tender-

Correspondence to: F. Schm.trr, Ph.D., Faculty of Dentistry, University of Western Ontario, London, Ont. N6A X2, Canada.

ness of masticatory muscles’ [30]. The purpose of this study is to examine the degree to which TMJPD patients differ from other pain patients and healthy controls along the dimensions of per- sonality type, response to illness, attitudes towards health care, and ways of coping with stress.

Krantz and Hedges [16] elaborate on 3 ways in which traits may contribute to any disease pro- cess. First, traits may be viewed as factors which predispose the individual to developing an illness. This theory originated with Dunbar [5] and Alexander [l] and was later applied to TMJPD by Moulton [25] and others [11,17]. It argues that

0304-3959/90,/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)

when conflict occurs in the individual’s life, which is not resolvable because of the underlying per-

sonality structure, the unresolved conflict results

in state changes which, in turn, lead to physiologi- cal changes. At this point, a feedback loop may be

established between the onset of the illness and the state responses to the illness thereby further

compromising the well-being of the individual. Given this model, which is essentially a psycho-

somatic one, TMJPD patients might be expected to differ along sundry dimensions of trait per-

sonality from healthy individuals or from other

pain patients whose pain is the result of trauma. The model, however, is less able to address the

issue of state changes, as elevations on these scales may reflect the unresolved conflict, be a result of the disease process, or be a function of both.

A second way in which traits may play a role in

the disease process is through coping styles. Here the emphasis is not so much on personality char- acteristics as it is on the ways in which individuals

cope with stressful situations. Coping strategies may be used to alter the relationship between the

individual and the environment or to control or minimize the stressful emotions or physiological

arousal that have been elicited. Presumably, inef- fective coping strategies result in adverse physio-

logical and behavioural changes which have conse- quences for the individual’s health [16].

A third way in which traits may contribute to disease has more to do with the end process of the

disease than with factors associated with its onset. This is often referred to as illness behaviour [16,19]. It is estimated that only 5% of those with TMJPD signs and symptoms actually seek treatment [19,28]. Although a change in symptom severity

may account for some of this. other factors may be responsible as well. These factors may include the degree to which the individual attends to his symptoms, their perceived meaning, attitudes to- wards health care, and accessibility to health care.

When the potential contribution of personality factors in TMJPD is considered within the context of this framework, the question is not as simple as whether personality factors play a causal role in the development of TMJPD or are consequences of it. Rather, we are presented with a series of questions about the role of diverse personality

variables in the onset. maintenance, and treatment outcome of the dysfunction. While not all of these questions can be addressed within a correlational study, particularly questions pertaining to the role

of state variables as causal or resultant factors. it is possible to explore some of these relationships

and evaluate their relative importance in TMJPD. On the basis of this review. the following re-

search questions are proposed: First, what is the

nature of the relationship. if any, between the chronicity. pain intensity. and perceived severity of the TMJ dysfunction. as one set of variables.

and diverse facets of personality, as another set?

Second. does the same type of relationship occur among other non-TMJ pain-related disorders? Third, are there any general or specific differences in personality type or style between TMJPD pa- tients or TMJPD subgroups, non-TMJ pain pa- tients, and pain-free. healthy subjects? Fourth. if there are differences between TMJPD patients

and the control groups, do these differences de- crease in response to a positive treatment out- come?

Method

Sdjects

A total of 356 subjects were used in the study. The first group was comprised of 206 consecutive

referrals to the Facial Pain Clinic at the University of Western Ontario. Patients meeting the TMJPD criteria outlined by Eversole and Machado [6] were categorized as either myogenic facial pain (n = 42) internal derangement type I (n = 69). internal derangement type II (n = 85) or internal

derangement type III (n = 6). An additional 3 patients were diagnosed as degenerative joint dis-

ease and 1 as symptomless click. The mean age for the dental group was 28.3 years (S.D. = 10.6) with 87.4% of the subjects female. The mean age and sex ratio are comparable to those reported

elsewhere [2,12,29]. The second group, a pain control group (n =

79) was comprised of consecutive referrals attend- ing outpatient physiotherapy clinics for the treat- ment of painful conditions arising from knee in-

155

juries (13) tend&is (8) bursitis (6) fractures

and sprains (5) surgery (3), shoulder injuries (11) whiplash and other neck injuries (7), back pain (12), and other sports injuries, accident, and minor injuries. Patients were selected for inclusion in the study by their physiotherapist with the proviso that the patients were not being treated for TMJPD or disorders of unknown, organic aetiology. The

mean age was 33.0 years (S.D. = 11.1) and 69.2% of the subjects were female.

The third group was a pain-free, healthy con-

trol group (n = 71) comprised of upper year psy- chology students. The subjects were from a class

and volunteered to be in the study. Subjects with chronic disease and/or pain were excluded. The

mean age was 29.7 years (S.D. = 8.4) with 76.1% of these female.

Materials

Questionnaire

A 30 page questionnaire was developed for the study. In addition to items pertaining to the onset and chronicity of the presenting problems, and to

the age and sex of the subject, the questionnaire contained the following major measures:

(a) Temporomandibular joint pain and dysfunc-

tion index (dysfunction index). The temporoman- dibular joint pain and dysfunction index was com-

prised of 10 items formatted on a 6-point Likert scale, anchored by ‘never’ and ‘always.’ These items, which dealt with such variables as pain, limited mobility, clicking, and grinding, were

culled from the dental literature and are symp- toms commonly reported by TMJPD patients. The

sum of these items, which can range from 0 to 50, was used to represent an overall subjective mea- sure of the involvement of the dysfunction. *

(b) Visual analogue scale (VAS). A numbered visual analogue scale was used to rate average pain intensity. Values ranged from 0 to 6 with the end-points anchored by ‘no pain at all’ and ‘as intense as I can possibly imagine.’

* A copy of this index is available from the authors upon

request.

(c) Basic Personality Inventory (BPI) [14]. The

BP1 is a lZscale, 240-item, true or false question- naire which measures components of psychopa- thology similar to those measured by the MMPI [13]. The BP1 was chosen over the MMPI because

the latter has been considered inappropriate for the use with pain patients [23,33]. The BP1 con- tains fewer pain-related items than the MMPI and its scales are relatively independent, with no item

overlap, allowing for greater discriminatory power. Each scale contains 20 separate items. The 12

scales include hypochondriasis, depression, denial, interpersonal problems, alienation, persecutory

ideas, anxiety, thinking disorder, impulse expres-

sion, social introversion, self-depreciation, and de-

viation. Four of these scales (alienation, persecu- tory ideas, thinking disorder, and deviation) were

not used in the present study as the subjects were

unlikely to exhibit marked signs of psychopa- thology.

(d) Multidimensional Health Locus of Control

(MHLC) [37]. The MHLC scale is a 3-dimen- sional, 18-item, Likert scale questionnaire desig- ned to measure the extent to which an individual

believes his health is or is not controlled or de- termined by his own behaviour [37].

(e) Illness Behavior Questionnaire (IBQ) [27].

The Illness Behavior Questionnaire is a 7-scale, 62-item, yes-no format questionnaire which mea-

sures various aspects of a patient’s attitudes and feelings towards his illness, as well as his percep-

tion of how others react to his illness, and his view

of his current psychosocial situation [34].

(f) Perceived Stress Scale (PPS) [4]. The Per- ceived Stress Scale is a 16item, 5-point Likert

scale questionnaire designed to measure the extent

to which situations in the subject’s life are per- ceived as stressful. A general measure of self-re- ported stress is obtained by summing across the 14 items.

(g) Ways of Coping (Revised). The Ways of Coping (Revised) scale [8,9] is an 8-factor, 66-item, 4-point Likert scale questionnaire designed to sample a wide variety of thoughts and acts people use to deal with stressful situations. In the present

study, the subjects were told that people have many different ways of coping with stressful events. They were instructed to select those

1%

strategies listed in the questionnaire that they might use in dealing with stressful situations.

Procedure

The 206 patients were seen by R.I.B. and di- agnosed according to the criteria outlined by

Eversole and Machado [6]. A diagnosis of myo- genie facial pain (MFP) was made in patients

whose main complaint was pain in the absence of

any joint sounds on opening and closing the

mouth. If limited opening was present, it could be attributed to myospasm and not to any structural

limitation in the joint. A diagnosis of internal derangement type I was made in subjects who experienced an opening click with or without a closing or reciprocal click. Deviation of the jaw on

opening did not preclude a diagnosis of internal derangement type I. The criterion for a diagnosis of internal derangement type II was met if sub-

jects reported a history of transitory closed lock. The patient’s jaw would occasionally lock while

opening; however, this difficulty could be readily

overcome if the patient put the mandible through lateral excursions or pushed manually on the joint. If the closed lock condition (less than a 25 mm

opening) could not be readily overcome and no clicking joint sounds were audible, a diagnosis of

internal derangement type Ill was made. Joint sounds may or may not be present in this condi-

tion [6]. Of the patients diagnosed as suffering from internal derangements types I, II, or III, 88.2% were experiencing pain at time of assess- ment; the others had experienced intermittent

pain. Following dental assessment and diagnosis by

R.I.B., the patients were asked to complete the questionnaires either in the clinic or at home. This

task took approximately 45 min. The patients were prescribed simple jaw exercises to conduct at home and were referred for ultrasound at a physi- otherapy clinic. These conservative and non-inva-

sive forms of treatment have been shown to be effective [2,32] and are generally used before more radical treatment is attempted.

At a 5 month follow-up, the TMJPD patients were contacted by mail and asked to evaluate their pain and symptoms and to complete and return a

questionnaire battery similar to that originally completed. Of those contacted, 99 returned the questionnaires. Responses were nearly evenly dis- tributed among those indicating that treatment

was not very effective. moderately effective, and very effective.

Results

Patient characteristics

The length of time the individual reported hav- ing the problem (chronicity) was based on a 7- point category scale with the lowest value repre- sented by ‘less than 1 month’ and the highest value by ‘more than 5 years.’ The length of time the TMJPD patients reported having the problem was slightly negatively skewed with 52.5% indicat-

ing that they had had the problem more than 1

year. The length of time that the pain controls reported the problem was more positively skewed with 67.1% reporting less than 12 months (Table

I). Clenching and grinding were considered pre-

sent if the subject indicated that he or she was aware of engaging in these behaviours at least

sometimes. In the present study, 58.7% of the

TABLE I

TMJPD, PAIN CONTROLS, AND PAIN-FREE, HEALTHY

CONTROLS WHO PARTICIPATED IN THE STUDY

Number

% female

Mean age

Chronicity

(mode in months)

TMJ dysfunction

index

Pain Intensity

TMJPD Pain Healthy

(SD.) controls controls

(S.D.) (SD.)

202 79 71

x7.1 6X.4 76.1

28.0 33.0 29.2

(10.4) (11.1) (X.4)

12-24 l-3

24.3 6.6 5.7

(7.7) (5.7) (5.7)

3.0 3.5

(1.7) (1.6)

TMJPD subjects reported that they clenched their teeth. This was significantly different (x*(2) = 9.42; P = 0.009) from that reported by the pain control sample (43.6%) and the healthy controls (40.6%). Whether or not clenching is indicative of stress, it would appear that it is more common for TMJPD patients to clench their teeth than it is for

other groups.

The number of TMJPD patients who actually attribute the onset of their problem to a specific

stressful event is relatively low (9.0%). It is almost as common for them to attribute it to dental work (6.5%) or an accident or illness (11.5%). Although

the number of patients who report that the condi- tion is aggravated by stress or worry, emotional upset, disturbed sleep, or certain situations is rela-

tively high (58.8%), it is comparable to and even slightly lower than that reported by the other pain patients (61.5%). It would seem that a painful

condition of any type is aggravated by emotional upset, stress or worry, and certain other situations.

Analyses

In order to determine whether there is a direct relationship between the 3 symptom-related varia- bles (independent variable set) and the personal- ity-related variables (dependent variable sets)

157

among the TMJPD subjects (n = 202), 4 canonical

correlation analyses were performed using SPSSx [35]. A canonical correlation analysis is a method for examining the relationship between multiple

independent and dependent variables [26,36]. Canonical correlation is superior to multiple re-

gression in that relationships of greater complexity may be examined with fewer analyses than would

be necessary using multiple regression. An excel-

lent overview of this approach is provided by Tabachnick and Fidel1 [36].

In each of the 4 analyses, the independent set included the chronicity, pain intensity, and per-

ceived severity of the TMJ dysfunction, as mea- sured by the TMJ dysfunction index.

First set of analyses

Is there a direct relationship between the chronic-

ity, pain intensity, and perceived severity of the TMJ

dysfunction as one set of variables and diverse aspects

of personality as another set? In the first analysis of this set, the dependent variables included the Per- ceived Stress Scale and the Basic Personality In-

ventory. The PSS and BP1 were treated as a set in this analysis and further analyses because they

both tapped what might be considered a broad range of traditional personality dimensions. In the second analysis of this set, the dependent variable

TABLE II

SUMMARY OF 4 CANONICAL CORRELATION ANALYSES OBTAINED FROM DATA ON THE TMJPD GROUP (n = 202)

The independent variable set in each analysis contains the items pain intensity, chronicity, and the dysfunction index score.

Dependent Rc Rc

variable sets SQR

PSS and BP1 0.36 0.13

IBQ 0.34 0.12

MHLC 0.30 0.09

df Sig. Redund.

of I.V.

(27, 517.57) NS -

(21, 514.54) NS

(9,447.96) 0.016 0.04

Redund.

of D.V.

0.04

struct.

coeff.

I.V.

(> 0.30)

_

Intens (0.96)

struct.

coeff.

D.V.

(> 0.30)

_

Powerf

others (0.94)

Dysfunc (0.61) Chance (0.53)

Ways of Coping 0.32 0.10 (24, 519.76) NS - _ _

Rc = canonical correlation; Rc SQR = squared canonical correlation; df = degrees of freedom; Sig. = significance; Redund. of

I.V. = redundancy of the independent variable set; Redund. of D.V. = redundancy of the dependent variable set; Struct. coeff.

I.V. = structure coefficients of the independent variable set; Struct. coeff. D.V. = structure coefficients of the dependent variable set.

158

included the Illness Behavior Questionnaire. The third analysis included the Multidimensional

Health Locus of Control and the fourth analysis,

the Ways of Coping scale.

The results of these 4 analyses are presented in Table II. Using Wilks’ criterion, the only analysis

to obtain significance was that for the dependent MHLC set (F (9, 447.96) = 2.30; P = 0.016).

However, the canonical variates accounted for only

9% of the shared variance. The results of the canonical analyses were confirmed using the more

traditional multiple regression approach. Summary. The results of the canonical correla-

tion analyses on the TMJPD patients suggest that there is not a meaningful, direct relationship be-

tween diverse measures of personality and mea- sures of chronicity, pain intensity, and the per-

ceived severity of the dysfunction. That is, in- creases in one set of variables are not associated

with either increases or decreases in the other set

of variables.

The lack of significance would not appear to be a function of intercorrelated variables as the as- sumptions of multicollinearity and singularity were

not violated. Furthermore, the variables were blocked into groups during the analyses, as indi-

cated, in order to reduce the probability of highly

intercorrelated variables, in different personality tests, appearing in the analyses.

Second set of analyses Among patients with non-TMJ pain-related dis-

orders, is there a direct relationship between the

chronicity and pain intensity of the physically pain-

fur condition as one set of variables and diverse

aspects of personality as the other set? In order to answer this question, the first set of analyses was repeated using the 79 pain controls. The analyses were the same with the exception that the dysfunc- tion index score was excluded from the symptom set as the pain controls were not experiencing

TMJ-related problems. All of the analyses were significant with the exception of that for the Ways

of Coping scale (Table III). Using Wilks’ criterion, the canonical correla-

tion (0.54) for the PSS and BP1 was significant (F

(18, 136) = 1.96; P = 0.016) and accounted for 28.9% of the variance between the 2 canonical

TABLE III

SUMMARY OF THE 4 CANONICAL CORRELATION ANALYSES OBTAINED FROM DATA ON THE PAIN CONTROLS

(n = 79)

The independent variable set in each analysis contains the items pain intensity and chronicity.

Dependent

variable sets

PSS and BP1

Rc

0.54

Rc

SQR

0.29

4

(18, 136)

Sig.

0.016

Redund.

of I.V.

0.154

Redund.

of D.V.

0.058

StNCt.

coeff.

I.V.

( z 0.30)

Intens (0.91)

Chron (0.48)

StNCt.

coeff.

D.V.

(> 0.30)

PSS (0.84)

Hypo (0.81)

Depr (0.47)

Anxi (0.30)

IBQ 0.56 0.31 (14, 140) 0.006 0.161 0.096 Intens (0.96)

Chron (0.36)

DisCon (0.89)

AffDis (0.84)

Hypo (0.69)

MHLC 0.41 0.17 (6,148) 0.005 0.079 0.043 Intens (0.77) Powerf

Chron ( - 0.58) Others (0.80)

Ways of Coping 0.35 0.12 (16,138) NS - _ _ _

Rc = canonical correlation; Rc SQR = squared canonical correlation; dj= degrees of freedom; Sig. = significance; Redund. of I.V. = redundancy of the independent variable set; Redund. of D.V. = redundancy of the dependent variable set; Struct. Coeff.

I.V. = structure coefficients of the independent variable set; Struct. Coeff. D.V. = structure coefficients of the dependent variable set.

159

variates. The canonical variate of the dependent disease conviction (0.89), affective disturbance

set extracted 20.1% of the variance from the PSS (0.84), and h~~hon~asis (0.69) were the im- and BP1 variables while the canonical variate of portant variables in the dependent set. As in the the independent set extracted 53.1% of the vari- PSS and BP1 analysis, pain intensity (0.96) was ance from the 2 symptom-related items. Consider- more relevant than chronicity (0.36). The variance ing the redundancy index for the symptom set, the (17.1%) accounted for by the canonical variates of PSS and BP1 measures were able to account for the MHLC dependent set and symptom set was 15.4% of the variance in that set. Pain intensity considerably less than that accounted for by the and chronicity were able to account for only 5.8% PSS/BPI and IBQ, although it was still significant of the variance in the dependent set. (F (6, 148) = 3.25; P = 0.005).

With a structure coefficient cut-off score of 0.30 for interpretation, perceived stress (0.84) hy- pochondriasis (0.81), depression (0.47), and anxiety (0.30) tended to be the more important variables in the dependent set. Among the symptom set, pain intensity (0.91) was more relevant than chro- nicity (0.48). Hence, as pain intensity and chronic- ity increase there is a tendency for perceived stress, hypochondriasis, depression, and anxiety to in- crease as well.

The analyses were confirmed using stepwise multiple regression. The proportion of variance accounted for was slightly higher using the multi- ple regression approach but this tended to be the result of only 1 or 2 variables meeting the criteria for entry into the regression equation.

The canonical correlation (0.56) for the IBQ dependent set also obtained significance (F (14, 140) = 2.34; P > 0.006) and accounted for 30.8% of the variance between the canonical variates. The IBQ variables were able to account for 16.1% of the variance in the symptom set. Pain intensity and chronicity were able to account for only 9.6% of the variance in the dependent set. An examina- tion of the structure coefficients indicates that

Summary. The results of these analyses indicate a significant but weak relationship between the chronicity and pain intensity of non-TMJ painful conditions, on the one hand, and personality, par- ticularly perceived stress, h~~hond~~is, depres- sion, anxiety, and disease conviction, on the other.

Third set of analyses Are there any personality differences among

TMJPD subgroups, pain controis, and pain-free, healthy controls? In order to compare TMJPD patients with controls who do not, themselves, have significant TMJPD s~ptomatology, indi-

TABLE IV

TMJPD, PAIN CONTROLS, AND PAIN-FREE, HEALTHY CONTROLS WHO MET THE INCLUSION CRITERIA

Number

% female

Mean age

Chronicity (mode in

months)

TMJ dys. index

Pain

intensity

MFP (SD.)

42

85.1

29.0 (11.3)

12-24

20.5 (7.0)

3.2 (1.7)

Type I (SD.)

69

82.9

29.0 (10.3)

7-12

24.7 (7.8)

3.0 (1.8)

Type II (SD.)

85

89.4

26.7 (9.9)

12-24

25.7 (7.6)

3.0 (1.7)

Pain controls (SD.)

56

83.9

31.5 (10.4)

1-3

6.3 (5.0)

3.6 (1.7)

Healthy controls (S.D.)

52

84.6

29.3 (7.8)

4.5 (4.4)

viduals in the 2 control groups whose score on the additional 12 males were randomly excluded from TMJPD index exceeded a level 1 S.D. below the the pain control group and 8 from the health) mean of the TMJPD group were excluded from control group in order to balance the groups on the analyses. This involved 11 subjects in the pain sex. The TMJPD groups classified as internal de- group and 7 in the pain-free group. Four subjects rangement type III (n = 6), degenerative joint dis- in the pain-free, healthy control group who re- ease (n = 3), or symptomless click (n = 1) were ported chronic headaches were also excluded. An excluded from the analyses due to il~sufficient

TABLE V

MEAN PRE-TREATMENT SCORES ON THE PERSONALITY MEASURES FOR THE TMJPD GROUPS, PAIN CONTROLS.

AND PAIN-FREE, HEALTHY CONTROLS

MFP

(n = 42)

(SD.)

Type I

(il= 69)

(SD.)

Type II

in = 85)

(SD.)

Pain

controls

(n = 56)

(SD.)

Healthy

controls

(n = 52)

(SD.)

Basic Personnliry Invennrory Hyp~hond~asis 8.0 (3.9)

Depression 3.4 (2.8)

Denial 6.2 (2.2)

Interpersonal 7.6 (3.3)

Anxiety 8.1 (3.9)

Impulse control 6.0 (3.4)

Social introver 4.4 (2.9)

Self depreciation 1.6 (1.8)

Perceived Stress Siule PSS 22.3 (10.4)

Illness Behavior Questionnaire Hypochondriasis I.5 (1.6)

Disease conviction 1.4 (1.3)

Psych vs. somatic 1.8 (1.0)

Affective inhib 2.0 (1.9)

Affective disturb 2.2 (1.8)

Denial 2.4 (1.5)

Irritability 2.1 (1.1)

Multidimensional Health Locus of Control Internal HLC 25.2 (5.0)

Powerful others 16.0 (4.9)

Chance 16.5 (5.4)

Ways of Coping Confrontive 6.1 (2.7)

Distancing 6.6 (3.1)

SeIf-controlling 9.5 (3.3)

Seeking SW supp 8.8 (3.3)

Accepting respons 4.9 (2.1)

Escape avoidance 6.2 (4.0)

Planful prob sol 7.6 (2.3)

Positive reappr 9.4 (3.7)

6.5 (3.4)

2.7 (2.8)

6.3 (2.7)

8.6 (3.2)

7.6 (3.9)

6.5 (4.0)

4.3 (3.4)

1.4 (1.6)

22.4 (8.7)

1.5 (1.6)

1.3 (1.1)

1.8 (1.0)

2.1 (1.8)

1.5 (1.6)

2.5 (1.3)

1.9 (1.1)

25.7 (4.9)

15.8 (5.2)

16.9 (5.5)

6.9 (2.9)

6.4 (3.1)

9.7 (2.8)

9.1 (3.6)

5.0 (2.1)

6.3 (3.9)

7.6 (2.6) 9.5 (4.1)

6.9 (4.2) 4.0 (4.2) 5.8 (2.4) 8.0 (3.7) 7.7 (4.4) 6.7 (3.7) 4.X (3.6) 2.1 (2.2)

21.2 (8.7)

1.6 (2.1)

1.2(1.3)

I .7 (0.9)

2.3 (1.8)

1.6 (1.8)

2.7 (1.4)

2.1 (1.3)

25.3 (5.0)

16.0 (5.6)

17.6 (5.2)

5.9 (3.2)

6.6 (3.4)

9.6 (3.4)

8.4 (3.6)

5.1 (2.4)

6.7 (4.9)

7.3 (2.9)

8.7 (4.5)

7.3 (4.1)

3.2 (2.7)

6.3 (2.9)

7.9 (3.5)

6.9 (3.3)

6.3 (3.4)

4.5 (3.3)

1.2 (1.6)

22.3 (8.5)

1.7 (1.9)

1.7 (1.4)

1.4 (0.9)

2.1 (1.8)

1.5 (1.6)

2.7 (1.6)

2.0 (1.4)

25.3 (5.3)

16.5 (7.0)

17.7 (5.2)

5.9 (2.9)

6.0 (3.3)

8.9 (3.6)

8.7 (3.5)

4.1 (2.4)

6.0 (4.2)

7.0 (2.8)

8.5 (4.5)

3.2 (2.6) *

2.2t2.5) *

5.5 (2.9)

6.9 (2.9)

6.8 (3.6)

5.3 (3.0)

3.8 (2.9)

1.3 (1.4)

19.4 (8.4)

26.8 (4.6)

15.7 (5.3)

16.8 (5.1)

6.3 (2.6)

5.9 (2.4) 9.3 (2.8)

9.1 (3.4)

5.2 (2.3)

5.8 (3.9)

7.9 (2.0) 8.8 (3.7)

* Using multivariate analysis of variance, the subjects suffering from pain (3 TMJPD groups and pain controls) differed significantly

from the healthy controls on Hypochrondriasis (P < 0.001) and Depression (P = 0.005).

161

sample size. Table IV presents data on the remain-

ing 3 TMJPD groups and the 2 control groups used in the following analyses.

A between-groups multivariate analysis of vari- ance (MANOVA) was performed in order to de-

termine whether there was a significant difference between groups on any of the dependent mea-

sures. In the first MANOVA, 9 dependent varia- bles were entered, including the PSS and the 8

scales from the BPI. Mean values for the variables in each of the 5

groups appear in Table V. The healthy controls score lower on all but one of the PSS and BP1 measures. Employing Wilks’ criterion, the overall

test for a group effect was significant (F (36, 1088.5) = 2.53; P < 0.001).

The Roy-Bargman stepdown analysis indicates a significant effect for hypochondriasis (F (4, 297) = 11.80; P -C O.OOl), depression (F (4, 296) = 3.12; P = 0.015), and anxiety (F (4, 295) = 2.57; P = 0.038). The strength of association (n2) be-

tween each significant variable and the group ef- fect was strongest for hypochondriasis (13.7%), followed by a marginal association for depression (4.1%) and anxiety (3.4%).

Post hoc follow-up tests indicate that the

TMJPD groups and the pain control group dif- fered significantly from the healthy controls on hypochondriasis and depression. There were no significant differences between the TMJPD groups

and the pain controls or among any of the TMJPD subgroups.

A second between-groups MANOVA was con- ducted on the Illness Behavior Questionnaire (IBQ). As the healthy controls were not suffering from any physical problems, they did not com- plete the IBQ. The analysis therefore tests whether there are any differences among any of the 3 TMJPD groups and the pain controls. There were no significant differences between any of the groups.

Two additional between-groups MANOVA

were performed using the 3 Multidimensional Health Locus of Control scales and the 8 Ways of Coping scales as the dependent sets. The groups did not differ on any of these measures.

Summary. Although the effect is weak, TMJPD patients and pain controls differ significantly from

pain-free, healthy controls on the hypochondriasis

and depression scales of the BPI. There are no differences among TMJPD subgroups or between TMJPD patients and the pain controls on any of

the measures. What these results seem to suggest, particularly in light of the differences between the pain controls and the healthy controls, is that a

painful condition may result in minor elevations

on hypochondriasis and depression. A stronger

test of this hypothesis would be to see if these elevations decrease in response to treatment.

Fourth set of analyses

Do scale elevations on the various personality

measures decrease as a function of treatment.? In

order to determine whether elevations on the per- sonality measures would decrease in response to a

positive treatment outcome, subjects were selected from among the 99 follow-ups on the basis of their pain intensity and TMJ dysfunction index scores after treatment. Response to treatment was con- sidered positive if either follow-up pain intensity

or dysfunction index scores were less than their

respective baseline assessment scores. Eighty-four TMJPD subjects met either or both of these liberal criteria. The condition of the remaining 15 was

unchanged or worsened. The overall effect of treatment was significant

(F (2, 74) = 43.70; P < 0.001). The effect was

significant for both pain intensity (F (1, 75) = 70.42: P < 0.001). which decreased from a mean

of 3.0 to 1.2, and the TMJPD index (F (1, 74) =

9.25; P = 0.003), which decreased from 26.0 to

19.9. The strength of the association (n’) was strongest for pain intensity (48.4%) and consider-

ably less for the TMJPD index (11.1 W), indicating that improvement weights more heavily on changes in pain intensity than changes in overall TMJ

symptomatology.

Table VI presents the pre-treatment and post- treatment scores for the various personality mea-

sures for the improved TMJPD subjects who com- pleted the follow-up questionnaire. On many of the scales there is a decrease from assessment to follow-up. The overall decrease among the depen- dent measures of the BP1 and PSS was significant (F (9, 74) = 2.59; P = 0.012). Roy-Bargman stepdown F tests indicate that the decrease was

162

TABLE VI

MEAN PRE-TREATMENT AND POST-TREATMENT SCORES ON THE PERSONALITY MEASURES FOR THE TMJPD PATIENTS (n = 84) WHO ANSWERED THE FOL- LOW-UP QUESTIONNAIRE AND RESPONDED TO TREATMENT.

TMJPD

Pre Post (n = 84) (n = 84) (S.D.) (S.D.)

Bask Pe~o~a~i~ Inuentory

Hypochondriasis Depression Denial Interpersonal problems Anxiety Impulse control Social introversion Self-depreciation

4.9 (3.7) 6.1 (3.7) * 3.6 (3.8) 2.8 (3.0) * 5.8 (2.4) 6.0 (2.7) 8.0 (3.7) 8.3 (3.8) 7.9 (4.2) 7.5 (4.2) 6.4 (3.1) 5.9 (3.2) 4.8 (3.4) 4.6 (3.2) 2.0 (2.0) 1.6 (2.0)

Perceiued Stress Scale

PSS 22.2 (9.2) 21.1 (7.9)

Illness Behavior Qscestionnaire

H~~hon~~is 1.6 (1.9) 1.3 (1.8) Disease conviction 1.3 (1.2) 0.7 (1.0) * Psych vs. somatic 1.9 (1.0) 2.0 (0.8) Affective inhibition 2.5 (1.9) 2.5 (1.8) Affective disturbance 1.8 (1.7) 1.4 (1.6) * Denial 2.5 (1.5) 2.1 (1.5) * Irritability 2.2 (1.3) 1.7 (1.2) *

Mulridimensiunu~ Health LocLcr of Control Internal 26.2 (4.7) 25.9 (4.4) Powerful others 16.1 (5.7) 15.8 (5.3) Chance 17.5 (5.3) 17.7 (5.5)

Ways of Coping

Confrontive Distancing Self-controlling Seeking social support Accepting responsibility Escape avoidance Planful problem solving Positive reappraisal

6.2 (3.0) 6.0 (2.5)

6.9 (3.4) 6.6 (2.9)

10.0 (3.2) 9.7 (3.2)

8.9 (3.3) 9.0 (3.3)

4.9 (2.4) 4.9 (2.2) 6.4 (4.1) 6.1 (3.7)

7.6 (2.8) 7.4 (2.5)

9.2 (4.3) 9.0 (4.3)

for depression (11.4%) than for hypochondriasis (5.4%).

Scores tend to decrease on the IBQ as well (I; (7, 77) = 6.68; P < 0.001). Stepdown tests indicate that the decrease is significant for affective dis- turbance (F (1, 82) = 7.52; P = 0.007), irritability (F (1, 81) = 13.06; P = O&01), denial (F (1, 79f = 7.53; P = 0.007), and disease conviction (F (1, 77) = 7.42; P = 0.008). The strength of the associ- ations, in decreasing order of significance, are: irritabi~ty (13.9%), disease conviction (8.8%), de- nial (8.7%), and affective disturbance (8.4%). As might be expected, these decreases did not occur over the more stable measures of personality styles for the MHLC or the Ways of Coping scales.

Summary. There was a significant decrease on the hypochondriasis and depression scales of the BP1 and the irritability, disease conviction, denial, and affective disturbance scales of the IBQ among TMJPD patients who responded positively to treatment. There were no changes on the MHLC or the Ways of Coping scale. This outcome sup- ports the hypothesis that at least some of the psychological distress observed in these patients may be a function of the physical condition itself. It suggests, rather than demonstrates, this as psy- chological distress and the TMJPD symptomatol- ogy may be unrelated but moderated by a third variable. Conceivably, the patients who seek treat- ment for this dysfunction may also tend to be somewhat more distressed. If this is the case, a reduction in symptomatology may have a benefi- cial effect on the anxiety or depression, for exam- ple, without the symptoms being the cause of the anxiety or depression. In a sense, there might be one less thing to worry about.

Discussion

* Employing multivariate analysis of variance techniques sig- nificant pre-tr~tm~t and post-treatment differences were faund on these variables. The remaining variables were not significantly different from one another.

significant for hypochondriasis (F (1, 81) = 4.66; P = 0.034) and depression (F (1, 80) = 10.31; P = 0.002). The strength of the decrease was stronger

~eta~ionship between pemonafity and TMJPD

One of the more interesting findings in this study is the lack of a meaningful canonical corre- lation for the TMJPD patients between the chro- nicity, pain intensity, and perceived severity of the TMJ dysfunction, as one set of variables, and diverse measures of personality, as the other set.

163

Increases in physical distress do not appear to be linearly related either to increases or decreases in psychological disturbance. This finding is interest- ing in light of some of the assumptions that have been made about the nature of the dysfunction and psy~holo~c~ disturbance. This finding is per- haps even more interesting when contrasted with the canonical correlation analyses for the pain controls. Three of the 4 analyses were significant and accounted for small but meaningful propor- tions of the shared variance.

The results of the canonical analyses suggest that it may be inappropriate to assume that differ- ent pain disorders are subsets of a larger, more all-encompassing pain syndrome. Many ap- proaches to the evaluation and treatment of pain patients seem to view them as members of a homogeneous population, assuming that psycho- logical responses to their disorder will be similar. In examining the relationship between personality and pain, it is ins~fi~ient to consider only the intensity and chronicity of the pain. The meaning or consequence of the pain may be of paramount importance. This needs to be followed up in future studies.

The effect of TMJPD on the well-being of the individual

The results of this study indicate that TMJPD patients and patients experiencing pain attributa- ble to an organic pathology score significantly higher on the h~~hond~asis and depression scales of the BP1 than do pain-free, healthy sub- jects. Furthermore, elevations on these scales de- crease in response to a positive treatment out- come. TMJPD patients do not seem to differ from either pain patients or healthy controls in their response to illness or in the ways in which they cope with stress. Even when the TMJPD patients are subdivided along a severity continuum, there still do not appear to be any major psychological differences between the TMJPD patients and the other groups. These results are in agreement with more recent studies in this area [10,15,18-22,341. In general, TMJPD patients are found to differ on diverse psychological measures from pain-free controls but not from other pain patients. Merskey

et al. [24] suggest that TMJPD patients may even have a lower rate of psychological illness, as de- tected by questionnaire, than other chronic pain populations.

The only other study [7] to have used a classifi- cation of TMJPD similar to the present one found small but significant differences on the MMPI between myofascial pain (MFP), atypical facial pain (AFP), and TMJ internal derangement pa- tients (TMJID), with the MFP and AFP groups scoring higher than the TMJID patients on hypo- chondriasis, depression, hysteria, psychopathic de- viate, psychasthenia, and schizophrenia. However, as there was no apparent control for the multiple ANOVAs computed, the actual number of signifi- cant differences between groups may be somewhat lower. Furthermore, as in the present study, the elevations were still within the normal range.

Scudds et al. [31] suggest that small elevations on the BP1 may reflect symptomatology rather than psychopa~olo~. An ex~nation of the BP1 scales indicates that TMJPD patients may be more likely to respond to certain items as a result of their presenting symptoms than other patients. This criticism has already been raised with respect to the MMPI by Merskey et al. [23] and Smythe

]331* When one considers that TMJPD patients ap-

pear to be relatively normal and do not seem to attribute the onset of the problem to stressful events, it is somewhat puzzling as to why TMJPD has been considered a psychosomatic disorder. This may, in part, be a function of the way in which health care professionals have interacted with the less manageable patients. As in any large sample of people, there are likely to be individuals with psychological problems additional to their presenting problems. Salter et al. [30] estimate that between 20% and 30% of patients consult their physicians for primarily psycholo~c~ reasons. If these patients are the ones who draw most of the attention then an association between psychopa- thology and the physical condition is likely to be made, especially if an organic pathology is not readily observable. Some support for this is offered through the cluster analytic work of Butterworth and Deardorff [3] who found that up to 26% of TMJPD patients may have severe emotional dis-

164

tress to the point of appearing psychopathological on screening measures. Generalizations about the

role of psychological factors in TMJPD may. in

part, be based on dentists’ and physicians’ interac- tions with these problematic patients.

Acknowledgements

This study was supported, in part, by an Ontario Graduate Scholarship to Robert F. Schnurr, an Endowment for Dental Research award to Ralph

I. Brooke, and Grant AO-392 from the Natural Sciences and Engineering Research Council of

Canada to Gary EL Roilman.

References 19

Alexander, F., Psychosomatic Medicine, Norton, New York.

1950.

Brooke, R.I., Stenn, P.G. and MothersiB, K.J., The diagno-

sis and conservative treatment of myofascial pain dysfunc-

tion syndrome, Oral Surg. Oral Med. Oral Pathol., 44

(1977) 844-852.

Butterworth, J.C. and Deardorff, W.W., Psychometric pro-

files of craniomandibular pain patients: identifying specific

subgroups, J. Craniomandib. Pratt., 5 (1987) 225-232.

Cohen, S.. Kamarck, T. and Mermelstein, R., A global

measure of perceived stress, J. Hlth Sot. Behav.. 24 (1983)

385-396.

Dunbar. F., Psychosomatic Diagnosis. Hoeber. New York.

1943.

Eversole, L.R. and Machado, L., Temporomandibular joint

internal derangements and associated neuromuscular dis-

orders. J. Am. Dent. Ass., 110 (1985) 69-79.

Eversole, L.R., Stone, C.E., Matheson, D. and Kaplan, H..

Psychometric profiles and facial pain, Oral Surg. Oral Med.

Oral Pathol., 60 (1985) 269-274.

Folkman, S. and Lazarus, R.. If it changes it must be a

process: study of emotion and coping during three stages of

a college examination, J. Pers. Sot. Psychol.. 48 (1985)

150-l 70.

Folkman, S., Lazarus, R., Dunkel-Schetter, C., DeLongis,

A. and Gruen, R., The dynamics of a stressful encounter:

cognitive appraisal, coping, and encounter outcomes. J.

Pers. Sot. Psychol., 50 (1986) 992-1003. Gale, E.N., Psychological characteristics of long-term

female temporomandibular joint pain patients, J. Dent.

Res., 57 (1978) 481-483.

Heiberg, A.N., Heloe, B. and Krogstad, B.S., The myofas-

cial pain dysfunction: dental symptoms and psychological

12

13

14

15

16

17

18

20

21

22

23

24

25

26

27

28

29

~tnd muscular functron. .4 preluninary study hv team ,q-

preach. Psychother. Psychosom.. 30 (197X) 8 I -Y:

Helkimo. M.. Epidemiological surveys of dysfunctmn of the

masticatory system. Oral Sci. Rev.. 7 (1976) 54 -h9.

Holden. R.. Item Subtlety, Face Validity and the Struc-

tured Assessment of Pcychopathology. Disserratmn, t_‘nt-

veraity of Western Ontario.

Jackson. I>.. The Basic Personality Inventor>. .4uthor.

London, On t.. 1976.

Keefe, F.J. and Dolan. E., Pain behavior and pain copmg

strategies in low back pain and myofascial pain dysfunction

syndrome patients. Pain. 24 (1986) 49-56.

Krantz, D.S. and Hedges. SM.. Some cautions for research

on personality and health, .I. Pers., 55 (1987) 351.-357.

Lefer. L.A.. A psychoanalytic view of a dental phenome-

non: psychosomatics of the temporomandibular joint parn-

dysfunction syndrome. Contemp. Psychoanal.. 2 (19661

135-141.

Malow. R.M., Grimm. L. and Olson, R.E.. Differences in

parn perception between myofascial pain dysfunction pa-

tients and normal subjects: a signal detection analysis, J.

Psychosom. Res., 24 (1980) 303-309.

Marbach, J.J. and Lipton. J.A.. Aspects of illness behavior

in patients with facial pain, J. Am. Dent. Ass.. 46 (197X)

630-638.

Marbach, J.J., Lipton, J-A.. Lund, P.R.. Delahanty, F. and

Blank, R.T., Facial pains and anxiety levels: considerations

for treatment, J. Prosthet. Dent., 40 (lY78) 434-437.

Marbach. J.J. and Lund. P., Depression. anhedonia and

anxiety in temporomandibular joint and other facial pain

syndromes, Pain, 11 (1981) 7% X4.

Marbach. J.J., Richhn. D.M. and Lipton, J.A., illness be-

havior, depression and anhedonia in myofascial face and

back pain patients. Psychother. Psychosam., 39 (1983) 47

54.

Merskey. H., Brown, J., Malhotra, L., Morrison, D. and

Ripley, C.. Psychological normality and abnormality in

persistent headache patients, Pain, 23 (1985) 35547.

Merskey. FL, Lau, C.L., Russell, ES., Brooke. R.I.. James,

M.. Lappano, S., Neilsen, J. and Tilsworth, R.H., Screening

for psychiatric morbidity. The pattern of psychological

illness and premorbid characteristics in four chronic pain

populations, Pain, 30 (1987) 141-157.

Moulton, R.E., Oral and dental manifestations of anxiety.

Psychiatry, 18 (1955) 261-273.

Pedhazur, E.J.. Multiple Regression in Behavioral Research

(2nd edn), Holt, Rinehart and Winston, New York, 1982.

Pilowsky, I. and Spence, N.D.. Patterns of illness behaviour in patients with intractable pain, J. Psychosom. Res.. 19

{19’75) 279-287.

Rugh, J.D. and Solberg, W.K.. Oral health status in the

United States: temporomandibular disorders. J. Dent.

Educ., 49 (1985) 398-405.

Salter. M., Brooke, R.I. and Merskey, H., Temporom~dib- ular pain and dysfunction syndrome: the relationship of

clinical and psychological data to outcome, J. Behav. Med..

9 (1986) 97-109.

165

30 Salter, M., Brooke, RI., Merskey, H., Fichter, G.F. and

Kapusianyk, D.H., Is temporo-mandibular pain and dys-

function syndrome a disorder of the mind?, Fain, 17 (1983)

151-166.

31 Scudds, R.A., Rollman, G.B., Harth, M. and McCain, G.,

Pain perception and personality measures as discriminators

in the classification of fibrositis, J. Rheumatol., 14 (1987)

563-569.

32 Selby, A., Physiotherapy in the management of tem-

poromandibular disorders, Aust. Dent. J., 30 (1985) 273-

280.

33 Smythe, H.A., Problems with the MMPI, J. Rheumatol., 11

(1984) 417-418.

34 Speculand, B., Goss, A.N., Hughes, A., Spence, N.D. and

Pilowsky, I., Temporo-mandibular joint dysfunction: pain

and illness behaviour, Pain, 17 (1983) 139-150.

35 SPSSx, SPSSx User’s Guide (2nd edn), McGraw-Hill, New

York, 1986.

36 Tabachnick, B. and Fidell, L., Using Multivariate Statistics,

Harper and Row, New York, 1983.

37 Wallston, K.A., Wallston, B.S. and DeVellis, R., Develop-

ment of the mu~ti~mensional health locus of control

(MHLC) scales, Hlth Educ. Monogr., 6 (1978) 160-170.